The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process content from untrusted external websites and user-provided files.
Ingestion points: Public website URLs and external DESIGN.md files as specified in SKILL.md.
Boundary markers: Absent; there are no specific instructions to separate external data from system instructions or to treat external text as untrusted.
Capability inventory: The skill can write or modify local files (CSS, Tailwind, and Markdown) and draft outreach messages.
Sanitization: Absent; the skill does not specify procedures for validating or escaping content extracted from external sources before using it in code generation.
[EXTERNAL_DOWNLOADS]: The skill utilizes the external domain designmd.supply and refers to the context-dot-dev/designmd-supply project to facilitate design token extraction.
[COMMAND_EXECUTION]: The workflow describes a "local context-dot-dev/designmd-supply run," which indicates the execution of local command-line tools to process design data.

designmd-brand-kit